Generally, information is recorded on an optical disk at a certain normal velocity (line velocity or angular velocity), and the information is continuously recorded in a spiral track in the state where pieces of the information are arranged in a time series order. When the recording information is read out from such an optical disk so as to be reproduced, the read-out signal from the optical pickup is output as a reproducing signal in real time.
However, there is no guarantee that the recording information is always correctly read out. This is because there is a probability that a disk is damaged in a process of mastering or manufacturing the disk, or is damaged while the disk is in use. In the case where the injure exceeds a permissible limit, the recorded "0" or "1" of the recorded digital signal may be erroneously read out.
In the CD system, a CIRC (Cross Interleave Read-Solomon Code), is used as an error correcting code. The CIRC is a combination of a read-solomon code which is an error correcting code having high random error correcting capability, and means for converting a burst error (a long drop-out) into a random error by interleaving. The interleaving is effective to disperse the burst error caused by damage or the like of a disk to thereby improve the efficiency in correction. That is, erroneous data is dispersed by deinterleaving so that the erroneous data can be corrected on the basis of other correction data.
However, even if the CIRC method is used, sometimes correction becomes impossible when a large quantity of data is made erroneous by severe damage to the recording medium. If the erroneous data are output then noises are caused or shock sounds are generated from a speaker depending on the location of the error bits.
Then, it is desired to produce data approximate to the correct value even if the data is not perfectly correct. For this purpose, an erroneous data compensation circuit such as a preceding-value holding circuit in which a value immediately before erroneous data is repeatedly output or an average-value interpolation circuit in which an average value of correct data immediately before and after erroneous data is output as compensated data.
In the case where it is desirable to reduce the generation of noises or shock as much as possible, it is desirable that the read-out information per se, which is to be supplied to the erroneous-data compensation circuit, has the lowest possible error rate so that more improvement in the quality of information may be expected.